Seemingly random but consistent failure to obtain metadata

araneae
For some reason, I can't get the data from these three URLs:

http://journals.royalsociety.org/content/da7ymnybftvh9ev0?p=a37b4c3274d043baad754ec10007732c&pi=0

http://journals.royalsociety.org/content/7w7nuknh0qnvmeej/?p=067b7c710bf9402d857da6cb2147a6d3&pi=1

http://journals.royalsociety.org/content/e2d8mp0nlpw7ab1x/?p=c2c321600c4449d2a017ace5dbdec02a&pi=8

All papers from 2002 or earlier seem to have this problem. More recent papers are obtained easily.

What happens with this is that the "get paper" button in the URL field appears as normal. I click it, the typical red box appears, but nothing appears in it. After the usual time interval, the red box goes away. Nothing else happens.

Even after restarting, and trying both of these links multiple times, I couldn't get it to work. When I use the "create new item from current page" thing it works, but of course it doesn't interpret the meta data correctly and isn't as nice.
  • fbennett
    This happens when a site translator crashes or fails to capture an essential part of the citation data (the title, or one part of an author's name, if I recall correctly). Almost certainly, there is some subtle difference in the formatting of these pages that causes the fault. The handy tool for tracing the cause of failures like this is Scaffold. Unfortunately, it only runs with Zotero 1.0, and I only have 1.5 on my desk at the moment. But somebody else may come along with news of where the trouble is. From there it's a short step to a fix.
  • fbennett
    What the heck. I had a few minutes before the start of faculty meeting, so I set up 1.0 and Scaffold and took a look. It looks like the scraper grabs RIS data from the page or a linked file, and parses that. This is generally a good idea, but two of the author (AU) fields have only a comma in them. This is stripped by Zotero, so you end up with two empty author fields. Everthing is being captured (including the author data), but those two empty fields cause the Zotero database save routines to barf.

    This will be easy to fix when I get a sec. But it might be that the fix should be applied to the RIS translator, which is used by lots of sites -- author names consisting only of a single comma are never going to make sense anyway, so maybe they should always to culled. Maybe Dan or someone else from inside the team can chime in on this? Looks like a quick one to solve.
  • fbennett
    @araneae,

    I've spent a couple of minutes this evening looking at the translator problem you've been having on the Royal Society Publishing site. A small fix in the RIS translator has gotten the site working correctly on my machine, but it will need to find its way into the Zotero distribution channel before it will reach your browser. If you're interested or if it's useful, I've put the amended code on the Zotero development discussion site, under the name "fixed_ris.sql". The link to the file area is here:

    http://groups.google.com/group/zotero-dev/files

    The rest is up to the team.

    Frank Bennett

This is an old discussion that has not been active in a long time. Instead of commenting here, you should start a new discussion. If you think the content of this discussion is still relevant, you can link to it from your new discussion.